Drinking during pregnancy: Investigating slippages between scientific data and the interpretation, communication and impact

This work was funded by the Social Sciences Faculty Research Fund at the University of Kent, as a collaboration between Robbie Sutton (Psychology) and Ellie Lee (SSPSSR) at the University of Kent, working with Bonnie Hartley as RA.

Background

Women are commonly, and increasingly, advised to abstain altogether from alcohol whilst pregnant or planning a pregnancy. Nonetheless, systematic reviews of scientific studies have been published that fail to detect adverse effects of fetal alcohol exposure (after the first trimester) on children’s subsequent development (e.g., Henderson et al., 2007). Several studies have observed a positive relationship between mild or moderate drinking during pregnancy and children’s subsequent developmental outcomes (Alati et al., 2008; Kelly et al., 2007; 2012; O’Callaghan et al., 2007). In a burgeoning, cross-disciplinary literature, many researchers have explored how this disconnection of policy and advice from scientific evidence arises from political and social-psychological factors such as lobbying, paranoid attitudes to risk, and sexism (e.g., Bell et al., 2015; Gavaghan, 2009; Lee, 2007a; Sutton et al., 2011). Similar disconnections are apparent regarding other issues such as maternal exercise, diet, breastfeeding, and parenting practices (e.g., Wolf, 2010).

The Research

In this interdisciplinary project, we asked whether the critical problem is merely the faulty translation of scientific evidence to policy, or has also to do with questionable practices that may distort the scientific evidence available to governments, charities and the public. Specifically, we employed mixed methods to examine how social-psychological and political factors may impinge on the analysis, interpretation and reportage of findings in scientific papers. We also critically examined how scientists, and subsequently journalists, publicly communicate scientific findings.

Our research examined a much-cited Plos ONE paper (Lewis et al., 2012) as a case study of these processes. Lewis et al. found that at age 8, children of mothers who drank during pregnancy had a significant IQ advantage over children of abstinent mothers. The apparent “benefit” of maternal drinking remained, but shrank, for children genetically disposed to metabolise alcohol slowly. Nonetheless, in the article and a press release, Lewis et al. interpreted their results as evidence that maternal drinking reduces children’s IQ. The widespread media coverage generally appeared to follow suit. Our research critically examined (1) the statistics and inferences that led Lewis et al. (2012) to their conclusion, despite the prima facie findings; (2) biases in the communication of the research by science journalists and by the researchers themselves; (3) the political and social-psychological factors that influence how people, conceivably including scientists themselves, prefer to interpret and communicate scientific findings of this sort.

Outcomes

The findings were presented and discussed as part of the event Policing Pregnancy: A one-day conference on maternal autonomy, risk and responsibility (April 2016, Royal College of Obstetricians and Gynaecologists).
See abstract and PPT.

Findings are discussed in this article:

Lee, E., Sutton, R. and Hartley B. 2016. ‘From scientific article to press release to media coverage: advocating alcohol abstinence and democratising risk in a story about alcohol and pregnancy’. Health, Risk and Society 18(5-6): 247-269.

Abstract

In this article, we follow the approach taken by Riesch and Spiegalhalter in “Careless pork costs lives’: Risk stories from science to press release to media’ published in this journal, and offer an assessment of one example of a ‘risk story’. Using content and thematic qualitative analysis, we consider how the findings of an article ‘Fetal Alcohol Exposure and IQ at Age 8: Evidence from a Population-Based Birth-Cohort Study’ were framed in the article itself, the associated press release, and the subsequent extensive media coverage.

We contextualise this consideration of a risk story by discussing a body of work that critically engages with the development and global proliferation of efforts to advocate for alcohol abstinence to pregnant (and pre-pregnant) women. This work considers the ‘democratisation’ of risk, a term used to draw attention to the expansion of the definition of the problem of drinking in pregnancy to include any drinking and all women. We show here how this risk story contributed a new dimension to the democratisation of risk through claims that were made about uncertainty and certainty.

A central argument we make concerns the contribution of the researchers themselves (not just lobby groups or journalists) to this outcome. We conclude that the democratisation of risk was advanced in this case not simply through journalists exaggerating and misrepresenting research findings, but that communication to the press and the initial interpretation of findings played their part. We suggest that this risk story raises concerns about the accuracy of reporting of research findings, and about the communication of unwarrantedly worrying messages to pregnant women about drinking alcohol.